Note: Descriptions are shown in the official language in which they were submitted.
CA 02101280 2000-08-31
1
DEVICE FOR DETECTING FAILURE OF BATTERY CELLS
BACKGROUND OF THE INVENTION
The invention relates to a device for detecting failure of
battery cells.
Storage batteries, notably those used in Uninterruptible
Power Supplies, are generally made up of a number of cells
connected in series. Up to now detection of failure of one or
more battery cells has only been able to be performed by
monitoring the individual voltage at the terminals of each
cell. This involves measuring a large number of voltages and
requires the terminals of the different cells to be
accessible.
The object of the invention is to achieve a device for
detecting failure of one or more battery cells which is
simpler and does not present the above-mentioned drawbacks.
SUMMARY OF THE INVENTION
According to the invention, there is provided a device
for detecting failure of at least one battery cell of a
battery comprising a plurality of cells connected in series,
said device comprising:
voltage measuring means for measuring the voltage at
terminals of said battery during discharge of the battery;
computing means, connected to said voltage measuring
means, for computing a quantity representative of the second
derivative of said voltage over time;
CA 02101280 2000-08-31
2
comparing means, connected to said computing means, for
comparing said quantity with a preset, positive or nil,
threshold that is indicative of a change in sign of the
second derivative; and
failure indicating means, connected to said comparing means,
for indicating said failure when said quantity is greater
than said threshold.
According to the invention, there is also provided a
device for detecting failure of at least one battery cell of
a battery comprising a plurality of cells connected in
series, said device comprising:
voltage measuring means for measuring the voltage at
terminals of said battery during discharge of the battery;
computing means, connected to said voltage measuring
means, for computing a quantity representative of the second
derivative of said voltage over time;
comparing means, connected to said computing means, for
comparing said quantity with a preset, positive or nil,
threshold;
failure indicating means, connected to said comparing
means, for indicating said failure when said quantity is
greater than said threshold;
current measuring means for measuring discharge current
of the battery;
power computing means, connected to said voltage and
current measuring means, for computing power delivered by the
battery;
storing means, connected to aid power computing means,
for storing power delivered by the battery;
CA 02101280 2000-08-31
2a
difference computing means, connected to said power
computing means and to aid storing means, for computing the
power difference between two successive power values;
means, connected to said difference computing means, for
comparing said power difference with a preset power
threshold; and
inhibiting means for inhibiting said failure indicating means
if said power difference is greater than said power
threshold.
According to a preferred embodiment, the device stores three
successive values of the voltage at the terminals of the
battery, measured at preset time intervals, and computes from
these three values two successive values of a quantity
representative of the first derivative of the voltage. It
then computes the quantity representative of the second
derivative by computing the difference between the two
successive values of the quantity representative of the first
derivative.
BRIEF DESCRIPTION OF THE DRAWINGS
Other advantages and features will become more clearly
apparent from the following description of an illustrative
embodiment of the invention, given as a non-restrictive
example only and represented in the accompanying drawings in
which:
Figure 1 represents, in block diagram form, an installation
comprising a device according to the invention.
CA 02101280 2000-08-31
2b
Figure 2 represents a particular embodiment of a flow chart
for implementation of the invention.
Figures 3 and 4 represent respectively the variation curves,
versus time, of the voltage at the terminals of the battery
and of the voltages measured at the terminals of some of its
cells.
Figure 5 illustrates an additional phase of a particular
embodiment of the invention.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Figure 1 illustrates the application of a device 1 for
detecting failure of one or more battery cells to a battery 2
of a UPS. The UPS, of conventional type, represented
schematically in the figure, is supplied by an AC voltage
source 3 and comprises an AC-DC converter 4 serially
connected with a DC-AC converter, or inverter 5, and a load
6, the battery 2 being connected to the output of the
converter 4.
The device 1 comprises an electronic processing circuit 7,
~., ; t h
",.
~1fl~.28~
3
microprocessor, connected to a display device 8. The electronic
processing circuit 7 receives on input signals representative of
the voltage U at the terminals of the battery.
A failure of one or more battery cells results, when the battery
is discharging, in a break in the slope of the curve U(t)
representative of the voltage U versus time. Such a break in the
slope can be detected by computing the second derivative of the
curve U(t). In normal operation, the second derivative of the
voltage is always negative when the battery is in the course of
discharging. If the second derivative becomes positive in the
course of discharging, this is representative of failure of one
or more cells of the battery.
Figure 2 illustrates a preferred embodiment of a flow chart
which can be implemented by the microprocessor 7. In figure 2,
detection of failure begins by a phase F1 in which a first
measurement U3 is made by the microprocessor 7 of the voltage U
at the terminals of the battery 2. This phase Fl is followed by
a phase F2 consisting of a second measurement U2 of the voltage
U at the terminal s of the battery after a preset time interval
8t. Then the microprocessor computes (F3) the difference D2
between U2 and U3. The time interval Qt between two successive
measurements being preset and constant, the difference D2 is
representative of the derivative of the curve U(t).
The microprocessor then makes (F4) a third measurement U1 of the
voltage U at the terminals of the battery 2, the time interval
Qt separating the measurements of U1 and U2. Then it computes
(F5) the difference D1 between U1 and U2 and (F6) the difference
D between D1 and D2. The difference D is representative of the
second derivative of the curve U(t) constituted from the three
successive measurements U3, U2 and U1 of U made at constant
preset time intervals Lit and stored, for example in a RAM 9.
Then D is compared with a threshold, preferably zero, during a
210128
4
phase F7.
If D is 1 ower than the threshold ( output N of F7 ) , the battery
cells are considered to be in good working order. The second
measurement U2 replaces (F8) the value U3 in the memory 9 of the
processing circuit 7. Then (F9) the first measurement U1
replaces the value U2 in the memory 9 and the difference D1
replaces the difference D2. Thus, the last two measurements are
stored, along with their difference, and the microprocessor can
restart a new cyc'_e to determine the second derivative D at F4,
by a new measurement of U1, the time interval 4t separating two
successive measurements. As a non-restrictive example only, .CS,t
can be about 20 or 30s.
In this way, the second derivative is computed automatically at
each new measurement Ul.
If D is greater than the threshold (output Y of F7), a failure
is indicated for example (F10) by display on the display means
8.
Indication can naturally be achieved by any other suitable
means, either visual or acoustic, locally or remotely.
In the embodiment represented in figure 2, the threshold
representative of a failure is zero. To avoid certain spurious
alarms it may be preferable to choose a threshold slightly
greater than zero.
It can be shown experimental 1 y that the second derivative D is
well representative of a fai'_ure. As an example, figure 3
represents a discharge curve U(t) obtained experimentally at the
terminals of an inverter battery made up of 29 cell units.
The successive values of U, measured with a time interval ,L~t of
~1~~.~
1 minute, and the successive computed values of D1 and D are as
shown below.
t(min) U(V) D1 D
1 347.7665
2 351.5218 3.8
3 350.925 -0.6 -4.4
4 349.7962 -1.1 -0.5
5 348.2835 -1.5 -0.4
6 347.4759 -0.8 0.7
7 346.4733 -1.0 -0.2
8 345.6198 -0.9 0.1
9 344.5427 -1.1 -0.2
343.6921 -0.9 0.2
11 341.6315 -2.1 -1.2
12 340.0479 -1.6 0.5
13 339.2599 -0.8 0.8
14 338.4148 -0.8 -0.1
337.4351 -1.0 -0.1
16 336.2036 -1.2 -0.3
17 334.727 -1.5 -0.2
18 333.2822 -1.4 0.0
19 331.594 -1.7 -0.2
329.9797 -1.6 0.1
21 328.1715 -1.8 -0.2
22 326.8366 -1.3 0.5
Anomalies (D > 0) are observed at the following times .
6 - 8 - 10 - 12 - 13 - 18 and 20 mins .
Figure 4 represents the curves Ue(t) representing the variations
in the voltages at the terminals of 5 cell units during this
discharge. Significant voltage drops (Ue > lOV) can be observed
en at least one of the cell units at the following times . 5 -
11 - 12 - 17 - 19 and 20 mins. The voltages at the terminals of
....
~1fl~2~0
6
the remaining cell units, not represented, did not present any
significant drop.
In the event of a large modification of the power delivered by
the battery, computation of the second derivative can lead to
spurious indications of failure. It is therefore preferable to
inhibit indication in such a case. An additional inhibition
phase, as represented in figure 5, can be inserted in the flow
chart in figure 2, for example after phase F5 or F6. This
inhibition phase consists in measuring (F11) the current Ib
delivered by the battery, measured by a current sensor 10. Then
the microprocessor computes (F12) the power delivered P1 = Ullb,
and computes (F13) the difference d p between this power and the
power measured in the previous measuring cycle. Then Lip is
compared (F14) with a preset threshold A. If ~p is not greater
than the threshold (output N of F14), the microprocessor
continues normally according to the flow chart in figure 2 and
goes on to phase F7. If on the other hand a p is greater than
the threshold (output Y of F14), phases F7 and F10 are short-
circuited and the microprocessor goes directly to phase F8, thus
inhibiting any indication of failure. Phase F9 is then completed
by storing P1 in P2 for the next cycle.
The detection device described above may be combined with a
microprocessor-based device for determining the battery backup
time. Indeed, a device of this kind determines the backup time
of a battery from measurements of its voltage U and of the
current Ib delivered by the battery and from computation of the
power delivered. The existing flow chart then merely has to be
completed by a computation of the second derivative and its
comparison with a preset threshold to enable indication of
failure of one or more battery cells to be achieved.
